The growing adoption of Artificial Intelligence (AI) in public administration has largely been framed in terms of efficiency gains and decision-making enhancement. However, this perspective overlooks a deeper transformation concerning the reconfiguration of organizational knowledge processes. This study aims to examine how AI contributes to knowledge creation, interpretation, integration, and institutionalization within public administrations, and to identify the organizational and institutional conditions that enable or constrain these processes. By viewing AI not just as a technical tool but as a system that shapes how knowledge is created, interpreted, and used, this study aims to provide a deeper understanding of public organizations as knowledge-based systems.
Methodologically, the research adopts a qualitative meta-synthesis of interdisciplinary literature on AI in public administration. A structured search was conducted in the Scopus database, resulting in a corpus of 108 relevant studies. The analysis combines computational text analysis techniques—such as lexical mapping and heuristic scoring—with an interpretive thematic synthesis. This approach enables the systematic identification of recurring knowledge processes, enabling mechanisms, and contextual constraints shaping the role of AI in organizational knowledge dynamics.
The findings reveal that AI operates as a knowledge-enabling and knowledge-processing infrastructure through three interrelated processes. First, AI enhances knowledge creation by identifying patterns and transforming data into actionable insights. Second, it supports knowledge interpretation and integration by enabling interactive and dialogical processes between human judgment and algorithmic outputs. Third, AI facilitates the institutionalization of knowledge through feedback loops, continuous learning, and adaptive system updates. However, the effectiveness of these processes depends on key organizational conditions, including transparent and participatory governance, experimental leadership, supportive organizational cultures, and adequate technical and infrastructural capabilities. Conversely, rigid hierarchies and control-oriented approaches tend to limit AI’s role to procedural functions, constraining knowledge development.
From a theoretical perspective, the study contributes to knowledge management and public administration literature by integrating fragmented research streams into a unified framework that conceptualizes AI as a socio-technical knowledge infrastructure. It challenges traditional views of rational decision-making by highlighting the emergence of hybrid knowledge processes that combine human and algorithmic reasoning. Practically, the findings suggest that public organizations should design governance mechanisms that promote transparency, participation, and learning, invest in leadership and data literacy, and adopt flexible regulatory frameworks to fully leverage AI as a generative knowledge resource.